This chapter focuses on building better hallways, and allowing entities to spawn inside them. Nearest neighbor, straight-line, and hallway spawning are all covered.
The latest chapter in my roguelike tutorial is now live: fun with layers. In previous chapters, we've built a system for applying layered procedural generation, through the builder pattern. This chapter applies these techniques, adds a few new ones, and brings the total number of builder combinations we can apply with what we have to over 2,000!
Chapter 36 of the Roguelike Tutorial is live! It focuses on refactoring our map work thus far into a generic builder-based idiom, making it easy to extend map building with more layers in the future.
You might want to hand-craft certain rooms or features, and have the game sometimes include them in your randomly generated levels. This chapter will help you do just that. It also builds upon the concept of meta-builders - that is builders that take the output of other builders, and improve upon the map.
Sometimes, you want to add some hand-crafted content into the mix of your level. You might want to use a whole level as a set-piece, and carefully place everything. You might want to include a cool structure on a map. This chapter walks you through designing prefabs, and teaches you to place them as either whole maps or into existing maps.
You've probably seen Waveform Collapse (WFC), if you follow gamedev media. It exploded onto the scene, and started being used in everything from Caves of Qud to AAA titles. It's a complicated algorithm that takes one source image and makes something similar (but not the same). This chapter walks you through the algorithm, and modifies it to better fit the Roguelike genre - adjacency is influenced by tile connectivity to encourage useful maps, and it takes a map as a source - so it can make your existing map creation algorithms more interesting.
This chapter uses Voronoi diagrams to generate hive maps - not at all unlike the inside of an ants nest or bee hive. It discusses the use of different distance heuristics to adjust the shape of your result.
The DLA chapter introduced enforced symmetry and brush size for map generation, and this chapter takes it one step further - using it for other algorithms. Drunkard's Walk in particular can benefit from this system.
Despite the scary name, Diffusion-Limited Aggregation Maps (DLA) are actually relatively easy to implement - and can produce beautiful results. This chapter walks you through the basic algorithm, with tunables for where the particles start and the rules for how they carve your map. It introduces brush sizes and enforced symmetry for your map builders.
Mazes and labyrinths (optionally including a Minotaur) are a mainstay of D&D and Roguelikes. This chapter walks you through the creation of a perfect maze - one that is guaranteed to be solvable, and well-connected. It also talks about the pros and cons of including this type of map in your game.
The Drunkard's Walk algorithm makes maps by simulating a large number of tunnelers moving randomly through a level, carving out the map as they go. This chapter walks you through implementing the algorithm, along with settings and tunables to help you get the map type you want - while remaining random on each play-through.
Cellular Automata is a great algorithm family, taking chaos and making order. It produces nice, organic levels with high connectivity - ideal for creating the feel of a large natural cavern, or a forest level. This chapter walks you through the algorithm, its implementation, and how it can make fun levels.